The PubMed dataset consists of 19717 scientific publications from PubMed database pertaining to diabetes classified into one of three classes. The citation network consists of 44338 links. Each publication in the dataset is described by a TF/IDF weighted word vector from a dictionary which consists of 500 unique words.

Yearly citation totals from each year of the MEDLINE/PubMed Baseline referencing citations back to year 1781. These totals may increase over time for a particular year as new citations are added. For example, 25 citations were listed for the year 1800 in the 2018 MEDLINE/PubMed Baseline, while the 2019 Baseline includes 387 citations for that year.

PubMed is a free resource supporting the search and retrieval of biomedical and life sciences literature with the aim of improving health–both globally and personally. The PubMed database contains citations and abstracts of biomedical literature. It does not include full text journal articles; however, links to the full text are often present when available from other sources, such as the publisher's website or PubMed Central (PMC). See the PubMed User Guide for more information. https://pubmed.ncbi.nlm.nih.gov/help/

Hype - PubMed dataset Prepared by Apratim Mishra This dataset captures ‘Hype’ within biomedical abstracts sourced from PubMed. The selection chosen is ‘journal articles’ written in English, published between 1975 and 2019, totaling ~5.2 million. The classification relies on the presence of specific candidate ‘hype words’ and their abstract location. Therefore, each article (PMID) might have multiple instances in the dataset due to the presence of multiple hype words in different abstract sentences. The candidate hype words are 35 in count: 'major', 'novel', 'central', 'critical', 'essential', 'strongly', 'unique', 'promising', 'markedly', 'excellent', 'crucial', 'robust', 'importantly', 'prominent', 'dramatically', 'favorable', 'vital', 'surprisingly', 'remarkably', 'remarkable', 'definitive', 'pivotal', 'innovative', 'supportive', 'encouraging', 'unprecedented', 'enormous', 'exceptional', 'outstanding', 'noteworthy', 'creative', 'assuring', 'reassuring', 'spectacular', and 'hopeful’. This is version 2 of the dataset. Changes include: Added “Year” variable. Removed “Abstract length” variable. Modified variable information due to updated probabilistic model of hype. Number of hype words - 35 (updated from 36 based on revised findings). File 1: hype_dataset_final.tsv Primary dataset. It has the following columns: 1. PMID: represents unique article ID in PubMed 2. Year: Year of publication 3. Hype_word: Candidate hype word, such as ‘novel.’ 4. Sentence: Sentence in abstract containing the hype word. 5. Hype_percentile: Abstract relative position of hype word. 6. Hype_value: Propensity of hype based on the hype word, the sentence, and the abstract location. 7. Introduction: The ‘I’ component of the hype word based on IMRaD 8. Methods: The ‘M’ component of the hype word based on IMRaD 9. Results: The ‘R’ component of the hype word based on IMRaD 10. Discussion: The ‘D’ component of the hype word based on IMRaD File 2: hype_removed_phrases_final.tsv Secondary dataset with same columns as File 1. Hype in the primary dataset is based on excluding certain phrases that are rarely hype. The phrases that were removed are included in File 2 and modeled separately. Removed phrases: 1. Major: histocompatibility, component, protein, metabolite, complex, surgery 2. Novel: assay, mutation, antagonist, inhibitor, algorithm, technique, series, method, hybrid 3. Central: catheters, system, design, composite, catheter, pressure, thickness, compartment 4. Critical: compartment, micelle, temperature, incident, solution, ischemia, concentration, thinking, nurses, skills, analysis, review, appraisal, evaluation, values 5. Essential: medium, features, properties, opportunities, oil 6. Unique: model, amino 7. Robust: regression 8. Vital: capacity, signs, organs, status, structures, staining, rates, cells, information 9. Outstanding: questions, issues, question, questions, challenge, problems, problem, remains 10. Remarkable: properties 11. Definite: radiotherapy, surgery

PubMed comprises more than 26 million citations for biomedical literature from MEDLINE, life science journals, and online books. Citations may include links to full-text content from PubMed Central and publisher web sites.

Not all articles in PMC are available for text mining and other reuse, many have copyright protection, however articles in the PMC Open Access Subset are made available for download under a Creative Commons or similar license that generally allows more liberal redistribution and reuse than a traditional copyrighted work.

AnonymousNodeGAE/pubmed dataset hosted on Hugging Face and contributed by the HF Datasets community

This dataset gathers three types of pairs: Title-to-Abstract (Training: 22,811/Development: 2095/Test: 2095), Abstract-to-Conclusion and Future work (Training: 22,811/Development: 2095/Test: 2095), Conclusion and Future work-to-Title (Training: 15,902/Development: 2095/Test: 2095) from PubMed. Each pair contains a pair of input and output as well as the corresponding terms(from original KB and link prediction results).

The MEDLINE/PubMed Baseline Repository (MBR) provides access to each MEDLINE/PubMed Baseline snapshot starting with the 2002 MEDLINE Baseline. Each baseline contains a snapshot of MEDLINE citations in the state they were at a given moment in time without the MeSH vocabulary updates and other revisions that occur during the year. The baseline snapshot is created at the beginning of each new MeSH Indexing Year. The records included in the MEDLINE/PubMed Baseline databases represent a static view of the data at the time each baseline database was created.

gayanin/pubmed-abstracts-noised-with-kaggle-dist dataset hosted on Hugging Face and contributed by the HF Datasets community

deddyext/nbs-pubmed-qa dataset hosted on Hugging Face and contributed by the HF Datasets community

Dataset Summary

A daily-updated dataset of PubMed abstracts, collected via PubMed’s API and published on Hugging Face Datasets.Each snapshot is versioned by date (e.g., 2025-03-27) so users can track historical changes or use a consistent snapshot for reproducibility.

Updated daily Each version tagged by date Abstract-only dataset (no full text)

  Dataset Structure

Column Type Description

pmid string Unique PubMed identifier

abstract string Abstract text… See the full description on the dataset page: https://huggingface.co/datasets/uiyunkim-hub/pubmed-abstract.

A file containing all Min/Max Baseline Reports for 2005-2023 in their original format is available in the Attachments section below. A second file includes a separate set of reports, made available from 2002-2017, that did not include OLDMEDLINE records.

MEDLINE/PubMed annual statistical reports are based upon the data elements in the baseline versions of MEDLINE®/PubMed are available. For each year covered the reports include: total citations containing each element; total occurrences of each element; minimum/average/maximum occurrences of each element in a record; minimum/average/maximum length of a single element occurrence; average record size; and other statistical data describing the content and size of the elements.

PubMed is a service of the U.S. National Library of Medicine that includes citations from MEDLINE and other life science journals for biomedical articles back to the 1950s.

PKG21 CSV files.

Dataset Summary

First 10k rows of the scientific_papers["pubmed"] dataset. 10:1:1 split.

  Usage

from datasets import load_dataset

train_dataset = load_dataset("ronitHF/pubmed-10k", split="train") val_dataset = load_dataset("ronitHF/pubmed-10k", split="validation") test_dataset = load_dataset("ronitHF/pubmed-10k", split="test")

A file containing all Misc Baseline Reports for 2018-2023 in their original format is available in the Attachments section below.

MEDLINE/PubMed annual statistical reports are based upon the data elements in the baseline versions of MEDLINE®/PubMed are available. For each year covered the reports include: total citations containing each element; total occurrences of each element; minimum/average/maximum occurrences of each element in a record; minimum/average/maximum length of a single element occurrence; average record size; and other statistical data describing the content and size of the elements.

Dataset for Journal recommendation, includes title, abstract, keywords, and journal.

We extracted the journals and more information of:

Jiasheng Sheng. (2022). PubMed-OA-Extraction-dataset [Data set]. Zenodo. https://doi.org/10.5281/zenodo.6330817.

Dataset Components:

• data_pubmed_all: This dataset encompasses all articles, each containing the following columns: 'pubmed_id', 'title', 'keywords', 'journal', 'abstract', 'conclusions', 'methods', 'results', 'copyrights', 'doi', 'publication_date', 'authors', 'AKE_pubmed_id', 'AKE_pubmed_title', 'AKE_abstract', 'AKE_keywords', 'File_Name'.

• data_pubmed: To focus on recent and relevant publications, we have filtered this dataset to include articles published within the last five years, from January 1, 2018, to December 13, 2022—the latest date in the dataset. Additionally, we have exclusively retained journals with more than 200 published articles, resulting in 262,870 articles from 469 different journals.

• data_pubmed_train, data_pubmed_val, and data_pubmed_test: For machine learning and model development purposes, we have partitioned the 'data_pubmed' dataset into three subsets—training, validation, and test—using a random 60/20/20 split ratio. Notably, this division was performed on a per-journal basis, ensuring that each journal's articles are proportionally represented in the training (60%), validation (20%), and test (20%) sets. The resulting partitions consist of 157,540 articles in the training set, 52,571 articles in the validation set, and 52,759 articles in the test set.

PostgreSQL query to select the ten journals with the highest number of publications containing the MeSH term “Leukemia” [20] on the complete PubMed data set.

jAPC = Annual percent changes calculated by JoinpointRegression Analysis*APC is significantly different from zero when P < 0.05JoinPoint regression analysis of different APC trends.